Introduction

Cyanobacteria are ubiquitous prokaryotic microorganisms, which are unique in their ability to perform oxygenic photosynthesis. Cyanobacteria harness solar energy and convert it to chemical energy in the form of NADPH and ATP while simultaneously splitting water into O2, electrons, and protons (Knoll 2008). This chemical energy is utilized for CO2 assimilation and to power other metabolic processes. Apart from fixing atmospheric CO2, some strains of cyanobacteria also have the ability to fix molecular nitrogen (Bothe et al. 2010; Esteves-Ferreira et al. 2017). Fixation of N2 is a highly energy-demanding process that is catalyzed by the O2-sensitive enzyme, nitrogenase. Over the course of evolution, N2-fixing cyanobacteria have developed temporal and spatial strategies for keeping nitrogenase active in the presence of the O2-producing photosynthetic apparatus (Fay 1992). Filamentous heterocystous cyanobacteria utilize the aforementioned spatial strategy by confining photosynthesis to vegetative cells and N2 fixation to specialized heterocyst cells. When combined nitrogen is not available, some vegetative cells differentiate into heterocysts in a semi-regular pattern within the filaments. In the filament, fixed nitrogen is transported from the heterocysts to the vegetative cells in the form of glutamate, while vegetative cells supply carbon to heterocysts (Curatti et al. 2002), mainly in the form of sucrose (Nürnberg et al. 2015). The heterocysts are surrounded by a glycolipid layer, which is minimally permeable to O2. This and the lack of photosystem II (PS II)-dependent O2 evolution and enhanced respiratory activity within heterocysts create a micro-oxic condition, which is favorable to the efficient function of the nitrogenase enzyme (Newton 2007).

The H2 metabolism of N2-fixing cyanobacteria is driven by two major enzymes: nitrogenase, which fixes atmospheric N2 and produces H2 as a byproduct, and uptake hydrogenase (Hup), which recycles H2 produced by the nitrogenase enzyme (Tamagnini et al. 2007; Bothe et al. 2010). In addition to uptake hydrogenase, many N2-fixing cyanobacteria have bidirectional hydrogenase (Hox) (Tamagnini et al. 2002). However, the contribution of Hox hydrogenase to overall H2 photoproduction in N2-fixing heterocystous cyanobacteria is not significant (Carrieri et al. 2011), and therefore, its role will not be further discussed. In the absence of N2 or any other compatible substrate, nitrogenase catalyzes only the reduction of protons to H2 (Hinnemann and Norskov 2006). H2 photoproduction activity in cyanobacterial cultures can therefore be enhanced by exposing the cells to severe N-deficiency (Dutta et al. 2005) and increased CO2 supply (Leino et al. 2012).

Re-routing the energy spent for cell growth toward H2 production may be achieved by keeping the cyanobacterial cells at minimum growth rates. Immobilizing the N2-fixing cyanobacteria in Ca2+-alginate films and simultaneously exposing these films to a severe N-deficiency can minimize the growth rate and boost the H2 yield, as described by Leino et al. (2012). Moreover, immobilization of cells allows for an uniform distribution of a dense cell layer in comparison to suspension cultures (Kosourov and Seibert 2009; Kosourov et al. 2018), which in turn improves the efficiency of light energy utilization (Kosourov et al. 2017).

Even though entrapment of H2-producing cyanobacteria in thin layer improves the H2 yield, this condition together with N-deficiency causes an imbalance in the C/N ratio in the cells, leading to severe physiological stress especially in the excess of light and under high CO2 level (Kosourov et al. 2014). Glycogen is a metabolite of importance in cyanobacteria and their adaptation to changing nutrient flux. N-deprivation is known to impact glycogen accumulation in cells (Hai et al. 2001; Hasunuma et al. 2013).

Oxidative stress is likely experienced by cyanobacteria entrapped in films and may be even more prominent in comparison to suspension cultures. Similar to naturally occurring biofilms (Latifi et al. 2009), O2 may reach saturation levels in these Ca2+-alginate films due to the high rate of O2 released by photosynthesizing cells. Additionally, the compromised permeability of the polysaccharide matrix to O2 can also contribute to the accumulation of O2 within the films, leading to increased oxidative stress in cells over time. A recent study showed that algal and cyanobacterial cells entrapped within thin alginate films experienced oxidative stress under conditions favorable to H2 photoproduction and that the intensity of oxidative stress was light dependent (Kosourov et al. 2017).

We recently showed that the exposure of alginate-entrapped cyanobacterium Anabaena sp. PCC 7120 to the long-term N-deprivation condition for the improved H2 photoproduction yield adversely affected cell fitness, especially in the uptake hydrogenase deficient mutant (ΔhupL) (Kosourov et al. 2014). Cyanobacteria demonstrated decreased photosynthetic activity in the vegetative cells over time and noticeable visual changes in the pigment composition. Cyanobacterial carotenoids play important roles as accessory pigments in light harvesting (Stamatakis et al. 2014) and participate in photoprotection (Domonkos et al. 2013). Carotenoids are also known to affect membrane fluidity and form barriers to penetration for small molecules such as O2 (Berglund et al. 1999), thereby acting against oxidative stress. Importantly, the resistance of cyanobacterial PS II to photoinhibition may also depend on the diversity of carotenoids produced by the cells, since carotenoids are known to function in reactive oxygen species (ROS) scavenging as well as in the suppression of ROS production (Kusama et al. 2015). These qualities of carotenoids, and the inability of animals to synthesize them, have resulted in their promotion as nutraceuticals with antioxidant, immuno-modulating and anti-cancer activity (Ibañez and Cifuentes 2013), and a market totaling US$1.5 billion as of 2014 (BCC 2015).

This work reports on the different effects of long-term H2 photoproduction conditions on specific physiological responses of three strains of N2-fixing filamentous cyanobacteria entrapped in Ca2+-alginate films. Carbon allocation and carotenoid composition were targeted due to their significant roles in the adaptation strategies of cyanobacteria to environmental stress. A greater understanding of these adaptations and their impact on H2 production and carotenoid profiles are important for developing integrated production platforms for future bio-industrial applications.

Materials and methods

Strains and growth conditions

To examine long-term H2 photoproduction by immobilized N2-fixing heterocystous filamentous cyanobacteria, we evaluated the following three strains: (i) wild-type Calothrix sp. 336/3, (ii) the model strain Anabaena PCC 7120, and (iii) the uptake hydrogenase deficient mutant of Anabaena PCC 7120 (ΔhupL, Masukawa et al. 2002). Unfortunately, attempts at generating an uptake hydrogenase mutant of Calothrix have been unsuccessful thus far. Stock cultures were maintained at room temperature in 150 mL Erlenmeyer flasks containing 50 mL of Z8 medium (Kótai 1972) without combined nitrogen (Z8x); the ΔhupL medium was supplemented with 25 μg mL−1 spectinomycin. The flasks were illuminated from the top with cool-daylight fluorescence lamps (Lumilux T8 15W/865; Osram) with a light intensity of about 30 μmol photons m−2 s−1 photosynthetically active radiation (PAR). For immobilization, cells were grown from an OD750 of approximately 0.2 in 1000 mL Erlenmeyer flasks containing 500 mL Z8x medium and were bubbled with 0.2 μm filtered air (Acro 37TF). Growth was at 25 °C under continuous illumination from two sides with cool-daylight fluorescence lamps providing ~ 50 μmol photons m−2 s−1 PAR at the surface of the flasks. Antibiotics were not added to ΔhupL growth cultures.

Immobilization

Log-phase cultures were harvested by centrifugation at 3000×g for 5 min, washed once in Z8x medium, and pelleted again by centrifugation. Cells were then entrapped in Ca2+-alginate films according to the method of Kosourov and Seibert (2009) adapted for cyanobacteria (Leino et al. 2012). Templates were prepared using insect screen and Scotch-type tape and then washed with 70% ethanol and double-distilled water. One gram of pelleted cells was mixed with 1 mL of sterile 4% w/v alginate (#71238, Sigma) dissolved in Milli-Q water. This mixture was pipetted on the template and spread evenly by hand using a sterile 5 mL glass pipette. The alginate matrix was stabilized by spraying 50 mM CaCl2 solution. Finally, the Ca2+-alginate hydrogel films were cut into 3 × 1 cm strips and washed with sterile Milli-Q water to remove excess CaCl2.

Chlorophyll a determination

The chlorophyll a (Chl a) content was determined in randomly chosen triplicates of the freshly prepared Ca2+-alginate strips. The alginate matrices were solubilized in 50 mM Na-EDTA solution (pH 7.0), and filaments were pelleted and washed in Z8x medium by centrifugation. The resulting pellets were extracted with 90% methanol, and Chl a was measured spectrophotometrically at 665 nm using extinction coefficient given by Lichtenthaler (1987).

Hydrogen assay and measurement

Freshly prepared alginate strips were transferred into 23 mL glass vials containing 5 mL Z8x medium. Vials were tightly sealed with caps containing PTFE-coated rubber septa and were placed in a growth chamber at 25 °C under continuous illumination from the top (~ 70 μmol photons m−2 s−1, Philips Master TL-D T8 15W/865) with cool-daylight fluorescence lamps. Intermediate light intensity (70 μmol photons m−2 s−1) was used since high light could increase oxidative stress over time. To trigger efficient nitrogenase-driven H2 photoproduction in immobilized filaments, the vials were flushed with argon (Ar) and supplemented with 6% CO2. Every 3 days, the headspace of vials was replaced with either (i) Ar containing 6% CO2 (Ar + 6% CO2), (ii) ambient air (air), or (iii) ambient air containing 6% CO2 (air + 6% CO2), and samples were incubated for 16–20 h. Following this 16–20-h incubation, the headspace in all vials (conditions i, ii, and iii) was returned back to Ar + 6% CO2. This was performed to restore efficient H2 photoproduction in immobilized cyanobacteria. These conditions have been described previously (Kosourov et al. 2014).

For determination of H2 yields, 150 μL of gas was sampled from the headspace of vials using a gas-tight syringe (Hamilton) and injected into a gas chromatograph (Clarus 500, Perkin Elmer) equipped with a thermal conductivity detector and a molecular sieve 5 Å column (60/80 mesh). Argon was used as a carrier gas.

Sample collection

To investigate the changes induced in the films throughout H2 photoproduction experiment, triplicate samples were collected at the beginning (0 h) and end (450 h) of the H2 production assay experiment.

Determination of glycogen and other carbohydrates

Samples were treated with 50 mM Na-EDTA (pH = 7.0) to break down the alginate polymer, and cells were harvested by centrifugation. Pelleted cells were resuspended in 1 mL methanol (HPLC grade). After removing the methanol-extracted pigments by centrifugation, the remaining pellet was dried using a SpeedVac for 10 min at 43 °C (chamber heated) and mixed gently in 2 mL 100 mM acetate buffer (pH = 4.8) with a pipette. This mixture was transferred to sealed Hungate tubes and autoclaved for 20 min at 120 °C for solubilization of glycogen. After cooling, 1 mL acetate buffer (pH = 4.8) containing 4 units of freshly prepared amyloglucosidase (1011-51G-F, Sigma) was added and mixed well. Sealed tubes were incubated at 55 °C overnight. The following day, 100 μL of the supernatant was enzymatically treated with hexokinase, and absorbance of the solution was measured at 340 nm to quantify glucose according to the glucose hexokinase kit (DiaSys). The pellets remaining in Hungate tubes were used later for the determination of other polysaccharides.

The pellets collected after the amyloglucosidase reaction in the glycogen assay were resuspended in 1 mL 3% H2SO4 (v/v). This mixture was autoclaved in tightly sealed Hungate tubes at 120 °C for 40 min. The pH was adjusted to neutral with the addition of 2 M NaOH, and the volumes were equilibrated. This step was followed by the use of the glucose hexokinase kit (described above).

Detection of protein oxidation

Total protein samples were isolated as previously described (Pollari et al. 2011) with minor modifications. Briefly, cell pellets were washed twice with ice-cold STNE buffer (0.4 M sucrose, 10 mM Tris-HCl, pH = 8.0, 10 mM NaCl, 20 mM Na-EDTA, and 50 mM DTT), before and after transfer to a beater tube. A one-third volume of acid-washed glass beads (G1145-500G, Sigma) was added, and cells were broken by a Mini-BeadBeater-24 (BioSpec Products) in a cold room, under darkness. A cycle of beating for 1 min and then rapid cooling on ice for 1 min was repeated six times. The supernatant was collected to a fresh Eppendorf tube and centrifuged at 1500×g for 5 min at 4 °C to remove any unbroken cells. The collected supernatant was used as the total protein sample.

Samples for Western blotting were prepared according to the Oxyblot Protein Oxidation Detection Kit (Millipore). Total proteins were separated by 12% SDS-PAGE (Mini Gel Protean), and the separated proteins were transferred to Immobilon-PVDF membranes (Millipore). The chemiluminescent HRP substrate (Millipore) was used for the detection of carbonylated proteins. The membranes were stained with Coomassie Brilliant Blue (R-250, Bio-Rad) to verify equal loading and even transfer of samples.

Immunoblotting

The extracted total proteins were solubilized in Laemmli buffer supplemented with 10% β-mercaptoethanol and 6 M urea by incubating for 1 h at 55 °C. Proteins were separated with 12% SDS-PAGE containing 6 M urea and transferred to PVDF membranes by electroblotting. D1 N-terminal-specific and NifH reductase-specific (AS01021A, Agrisera) antibodies were applied to the membranes. The rabbit anti-hen IgY (H&L), alkaline phosphatase conjugate (Agrisera), and the CDP star chemiluminescence kit (New England Biolabs) were used for detection. Membranes were stained with Coomassie Brilliant Blue (R-250, Bio-Rad) to verify equal loading and even transfer of the samples.

Extraction and identification of carotenoids and tocopherols

Samples for HPLC analysis were collected and treated with 50 mM Na-EDTA (pH = 7.0) to break down the alginate and release cells. EDTA was removed by pelleting the cells and washing once with Z8x. Harvested cells were frozen in liquid nitrogen and stored at − 80 °C until processed. Pigments that are soluble in organic solvents were extracted from the pellets using 100% methanol in the dark at 10 °C overnight. This process was repeated until the complete extraction of the pigments. The solute was centrifuged and filtered through a 0.2-μm filter (Millipore) to remove any remaining debris. Filtered samples were analyzed by high-performance liquid chromatography (Agilent) using a C18-encapped column (LiChroCART 125-4, Merck). The pigments were eluted with two solvents, A and B, applied consecutively at a constant flow rate of 0.5 mL min−1. An isocratic run with solvent A (acetonitrile/methanol/0.1 M Tris-HCl, pH 8.0, 72:8:3 by vol.) for 4 min was followed by a linear gradient of solvent B (methanol/hexane, 4:1 by vol.) from 0 to 100% for 15 min. After the gradient, the isocratic run of solvent B was applied for another 26 min. Standards for the following pigments were run prior to running the samples: chlorophyll a, β-carotene, zeaxanthin, echinenone, canthaxanthin, myxoxanthophyll, and α-tocopherol (DHI LAB Products). The rest of the pigments were identified by LC-MS based on ion masses and spectral properties (Kosourov et al. 2016).

Results

Long-term hydrogen photoproduction by immobilized cyanobacteria

To study the physiological acclimation of N2-fixing heterocystous filamentous cyanobacteria to long-term H2 photoproduction conditions, we used three different treatments (detailed in the “Materials and methods” section): (i) the N-deprived H2 photoproduction condition of Ar + 6% CO2, (ii) periodic short-term treatments with air only, and (iii) periodic short-term treatments with air + 6% CO2. Treatments using air (thus adding N) were examined due to the demonstrated effect of added N on cell fitness and prolonged H2 photoproduction (Kosourov et al. 2014). Treatments were applied to (i) the native strain, Calothrix sp. 336/3; (ii) the model organism, Anabaena PCC 7120; and (iii) the uptake hydrogenase deficient mutant of Anabaena PCC 7120 (ΔhupL).

Treatments of cyanobacterial filaments entrapped within Ca2+-alginate films with air, or air + 6% CO2 significantly increased the cumulative H2 production yield of Calothrix sp. 336/3 cells compared to the N-deprived H2 photoproduction condition (Fig. 1a, 450 h). The effect of added N only became significant after around 250 h, when H2 production yield obtained in the Ar + 6% CO2 condition plateaued (Supp. Fig. S1a). This plateau has previously been linked to the declined photosynthetic efficiency at the conclusion of H2 photoproduction experiments (Kosourov et al. 2014). H2 photoproduction levels of the Anabaena sp. PCC 7120 were much lower than those of the other two strains (Fig. 1a). The wild-type Anabaena strain contains the [Ni-Fe] uptake hydrogenase (Hup) in heterocysts that recycles produced H2 very efficiently (Masukawa et al. 2002). In agreement with previous work (Kosourov et al. 2014), no significant changes were observed between treatments. At almost 0.45 mol of H2 m−2, the highest H2 yields were observed for the Hup-deficient (ΔhupL) mutant exposed to periodic air treatments (Fig. 1a, 450 h). Both conditions where CO2 was supplemented during H2 production and treatments (Ar + 6% CO2 and air + 6% CO2) resulted in plateaus of H2 production yield at around 150–175 h (Supp. Fig. S1c).

Fig. 1
figure 1

Long-term H2 photoproduction (a) and net O2 evolution (b) yield from Calothrix sp. 336/3, Anabaena sp. PCC 7120, and the ΔhupL mutant of Anabaena PCC 7120 entrapped in alginate films. The gas phase in the headspace of the vials was changed periodically with Ar supplemented with 6% CO2 (Ar + CO2), air (air), or air supplemented with 6% CO2 (air + CO2). The values are the mean of the three biological replicates (± SD, n = 3)

Full size image

Net O2 evolution yields were very similar between all strains over the three tested treatments (Fig. 1b), with the highest C/N condition (Ar + 6% CO2) resulting in the lowest O2 production for all strains. While oxygen production for Calothrix was generally in line with H2 production, ΔhupL demonstrated greater O2 (relative to H2) production under the air + 6% CO2 treatment.

Effects of the H2 photoproduction condition on the C/N balance of cyanobacteria: carbon allocation

During long-term H2 photoproduction, a significant boost in glycogen accumulation was observed for both strains of immobilized Anabaena, over all treatment conditions (Fig. 2). However, Calothrix sp. 336/3 did not show the same trend. The Ca2+-alginate-entrapped Calothrix cells accumulated significantly less glycogen than Anabaena under all studied conditions. An additional processing of samples to quantify total carbohydrates (excluding previously removed soluble sugars and glycogen) revealed large increases for the Calothrix strain, at levels clearly exceeding those of both Anabaena strains. The amount of total (non-glycogen) stored carbohydrates associated with the Calothrix cells increased over the course of the H2 photoproduction experiments and reached maximum levels at the conclusion of the air + 6% CO2 experiment at 450 h, where an approximately threefold increase was observed in comparison to the 0-h sample (Fig. 2). This increase easily surpassed final amounts of the two other treatment conditions, which is somewhat surprising given that the Ar + 6% CO2 treatment should induce the largest nitrogen deprivation relative to available carbon. While both Anabaena strains also demonstrated temporal increases in total (non-glycogen) stored carbohydrates, a large difference was not observed between treatment conditions in the wild-type strain. For ΔhupL, however, this increase was larger for the air-only treatment. Thus, the addition of CO2 (in either air or Ar) while inducing lower N relative to C did not induce a noticeable higher accumulation of C (glycogen and/or other stored carbohydrates). The apparent failure of Calothrix and ΔhupL to demonstrate maximum C storage under the highest C/N ratio conditions is likely due to the poor cell fitness for these strains toward the end of the long-term experiments employing these challenging conditions. This drop in cell fitness is reflected in plateaus of H2 production at around 250 h for Calothrix under Ar + 6% CO2 (Supp. Fig. S1a) and 175 h for ΔhupL under both Ar + 6% CO2 and air + 6% CO2 (Supp. Fig. S1c) and final (450 h) levels of Chl a (Supp. Fig. S2).

Fig. 2
figure 2

Glycogen and total (non-glycogen, stored) carbohydrates in Calothrix sp. 336/3, Anabaena sp. PCC 7120, and ΔhupL films. The samples were collected in the beginning of the experiment (0 h) and after 450 h under different treatment conditions. The values are the mean of the three biological replicates (± SD, n = 3), and the significance between 0 h (control film) and 450 h (Ar + 6% CO2, air, air + 6% CO2) was evaluated with Student’s t test (two-tailed; an asterisk represents: *p < 0.05; **p < 0.01; ***p < 0.001)

Full size image

While the wild-type Anabaena demonstrated roughly equivalent divisions of stored carbohydrates (between glycogen and non-glycogen, Fig. 2), the bias toward non-glycogen stored carbohydrates in Calothrix was obvious. Since Calothrix sp. 336/3 is a naturally benthic strain, the large levels of non-glycogen stored carbohydrates were likely stored outside the cells, as part of the carbohydrate-enriched extrapolymeric substances (EPS).

Oxidative stress experienced by cells under long-term H2 photoproduction conditions

In order to investigate the extent of oxidative damage in alginate-entrapped cyanobacteria, we performed immunoblotting for the detection of carbonylated proteins (an indicator of oxidative damage to protein) at 0 and 450 h. As is evident in Fig. 3, an increase in the protein oxidation of varying intensity was observed for all strains, under all tested treatments, compared to the 0 h samples. While the cells are at different physiological states at 0 and 450 h, alginate-entrapped cyanobacteria under different light intensities demonstrated a correlation between light intensity and protein oxidation (Kosourov et al. 2017). This indicates that entrapment and placement of cells under H2-photoproducing conditions contributes to the large increases in protein oxidation observed here.

Fig. 3
figure 3

Detection of carbonylation of proteins due to oxidative damage from total protein extracts of the three alginate-entrapped strains. The samples were collected in the beginning of the experiment (0 h) and after 450 h under different treatment conditions: (1) control, 0 h; (2) Ar + 6% CO2; (3) air; (4) air + 6% CO2

Full size image

We also evaluated the responses to oxidative stress by monitoring the accumulation of the lipid-soluble antioxidant α-tocopherol, which is known to efficiently scavenge singlet oxygen (1O2) in phototrophs (Rastogi et al. 2014). We detected notable increases in the accumulation of α-tocopherol for all strains under all tested conditions (Fig. 4), in order to combat lipid peroxidation (Steiger et al. 1999). In general, Calothrix samples showed approximately five- to sixfold higher α-tocopherol content than Anabaena strains, with temporal increases (at 450 h) being three to five times that of the 0 h sample. In Anabaena strains, the increase ranged from approximately 6- to 7-fold in the wild-type and 9- to 11-fold in ΔhupL. Overall, there was no obvious difference in the accumulation of α-tocopherol over the different treatments, although the Ar + 6% CO2 treated cells of Calothrix that showed a slightly lower accumulation of α-tocopherol than the air-treated cells.

Fig. 4
figure 4

α-tocopherol content in Calothrix sp. 336/3, Anabaena sp. PCC 7120, and the ΔhupL mutant at the beginning of the experiment (0 h) and end of the experiment (450 h). The values are the mean of the three biological replicates (± SD, n = 3), and the significance between 0 h and 450 h (Ar + 6% CO2, air, air + 6% CO2) was evaluated with Student’s t test (two-tailed; an asterisk represents: *p < 0.05; **p < 0.01; ***p < 0.001)

Full size image

Specifically, we were interested in possible oxidative damage to the PSII reaction center protein D1 and to nitrogenase. The D1 protein (encoded by psbA) is the primary target of light-induced oxidative damage in cyanobacteria. In order to evaluate the level of PSII in alginate-entrapped cyanobacteria, the PSII reaction center protein D1 was probed using a specific antibody for the N-terminal. By the end of the 450-h experiments, decreases were observed in D1 content for both Anabaena strains under the Ar + 6% CO2 condition (Fig. 5). The ΔhupL samples also demonstrated decreased D1 by 450 h under the air + 6% CO2 treatment. At the 450 h mark, levels of Calothrix sp. 336/3 D1 were quite stable across the three H2 production treatments investigated (Fig. 5). The relative stability of Calothrix D1 levels across the different treatments indicates that this strain may be able to undertake D1-recovery under adverse environmental conditions, such as the N deprivation studied here. This resilience of the Calothrix D1 protein also supports our previous demonstration of higher photosynthetic efficiency under the Ar + 6% CO2 condition, relative to both Anabaena strains (Kosourov et al. 2014). Importantly, the levels of D1 at 450 h (Fig. 5) correlated positively with cumulative H2 photoproduction yields (Fig. 1) both in Calothrix sp. 336/3 and ΔhupL strains.

Fig. 5
figure 5

Expression levels of targeted proteins (D1 and NifH) in Calothrix sp. 336/3, Anabaena sp. PCC 7120, and ΔhupL mutant. The samples were collected from (1) control films at 0 h; (2, 5) Ar + 6% CO2; (3, 6) air; and (4, 7) air + 6% CO2 at 250 h (2, 3, 4) and 450 h (5, 6, 7)

Full size image

NifH, the dinitrogenase reductase subunit of the nitrogenase enzyme (Bothe et al. 2010), was chosen as a marker to study the accumulation level of nitrogenase in heterocysts. While both Anabaena strains demonstrated increased accumulation of NifH by the end of the experiment, levels in Calothrix decreased under all treatments by the end of the experiment (450 h), most obviously, under the Ar + 6% CO2 treatment (Fig. 5). This result is interesting given the high H2 production levels of Calothrix at 450 h and previous demonstration of similar nitrogenase activity between the strains, albeit under short-term planktonic conditions (Leino et al. 2014). Thus, it is possible that the high H2 production of Calothrix is neither linked to the maximum capacities of the enzymes involved (Leino et al. 2014) nor the abundance of the nitrogenase enzyme at protein level.

Carotenoid profiles of immobilized cells under long-term H2 photoproduction conditions

Both our previous (e.g., Kosourov et al. 2014) and current research have demonstrated color changes in alginate-entrapped cells during long-term experiments (data not shown), indicative of the effects of changes in carotenoid content and Chl a. The strongest accumulation of total carotenoids was observed in Calothrix cells, where temporal increases were more than twofold for all treatments (Fig. 6). Temporal increases were also observed for both Anabaena strains, the largest of which were for the samples periodically treated with air. Total carotenoids also increased relative to Chl a in all strains by the end of the experiment (Fig. 6, squares). The most noticeable increase was for the ΔhupL strain when treatments were supplemented with CO2; this large relative increase was mostly due to the loss of Chl a in the cells (Supp. Fig. S2). Anabaena sp. PCC 7120 demonstrated the same trend as the ΔhupL strain, with final Chl a concentrations lower in both samples treated with supplemental CO2. Interestingly, Calothrix sp. 336/3 was the only strain to increase Chl a under all treatments by the end of the experiment. A similar experiment whereby Chl a was evaluated under Ar + 6% CO2, and samples taken at 340 and 600 h demonstrated decreased Chl a at the earlier time point and a later recovery (Kosourov et al. 2016), indicating that the increases are seen here (Fig. 6, Supp. Fig. S2) were likely preceded by an earlier decrease.

Fig. 6
figure 6

Total carotenoid content in Calothrix sp. 336/3, Anabaena sp. PCC 7120, and the ΔhupL mutant samples collected at the beginning of the experiment (0 h) and end of the experiment (450 h). The values are the mean of the three biological replicates (± SD, n = 3), and the significance between 0 h and 450 h (Ar + 6% CO2, air, air + 6% CO2) was evaluated with Student’s t test (two-tailed; an asterisk represents *p < 0.05; **p < 0.01; ***p < 0.001)

Full size image

Carotenoid profiles obtained at the beginning of experiments differed between the two Anabaena strains and Calothrix, more so than the response of the profiles to the three different treatment conditions (Table 1). While profiles differed between the strains, the amounts of the most abundant carotene in cyanobacteria, β-carotene, were relatively stable at around 20–30% for all strains and most treatment conditions, the only significant exception being a decrease to 18.3% for Calothrix under the air + 6% CO2 treatment condition. Under this treatment, Calothrix cells demonstrated a shift to cis β-carotenes, which are more efficient antioxidants than trans isoforms and are predicted to be localized closer to the reaction centers (Cerezo et al. 2012). Calothrix demonstrated a greater diversity of xanthophylls, which have been demonstrated to be stronger ROS scavengers than β-carotene (Steiger et al. 1999; Domonkos et al. 2013) and play a role in protecting cyanobacterial membranes from photo-oxidation (Zhu et al. 2010). Echinenone, a xanthophyll synthesized from β-carotene, was present in all strains, but levels in Calothrix were more than double those of Anabaena under the control growth condition (0 h, Table 1). All three strains demonstrated decreases in echinenone content, which were more pronounced for the treatments with the highest C/N ratios (Ar + 6% CO2 and air + 6% CO2). For Calothrix, there were small temporal increases in 3′-hydroxyechinenone, a derivative of echinenone and a known co-factor of orange carotenoid protein (OCP), which performs non-photochemical quenching (NPQ) in cyanobacteria (Maksimov et al. 2015). It should be noted that the changes in β-carotene, 3′-hydroxyechinenone, and zeaxanthin content of both Anabaena strains were not as pronounced as reported previously (Kosourov et al. 2016), probably due to a much lower light intensity applied to the films. Indeed, the levels of these carotenoids in Anabaena depend strongly on the light intensity (Kosourov et al. 2017).

Table 1 Carotenoid composition of Calothrix sp. 336/3, Anabaena sp. PCC 7120, and ΔhupL strains under diazotrophic growth condition (Z8x medium, +N2) and after long-term H2 photoproduction (Z8x medium, −N2). The experiments were performed with cyanobacteria entrapped in thin Ca2+-alginate films. The pigments were extracted in the beginning (0 h) and in the end (450 h) of the H2 photoproduction experiment
Full size table

The temporal decrease in echinenone led to the accumulation of alternative xanthophylls in all strains so that the overall xanthophyll content appeared relatively stable. Under both air treatments (air and air + 6% CO2) where Calothrix cells were most healthy, increases were observed in nostoxanthin, a carotenoid not detected in Anabaena cells, and in 2-hydroxymyxol 2′-methylpentoside and myxol 2′-methylpentoside. Since the glycoside moiety in the above-mentioned myxoxanthophylls from Calothrix has not been determined, it might be either fucoside, rhamnoside, or chinovoside. Similar increases were observed for 4-ketomyxol 2′-fucoside in both Anabaena strains, where content was more than doubled under all treatment conditions. The cells which suffered from the most bleaching and which were clearly unfit (D1 content, Fig. 5) were the ΔhupL cells under severe C/N stress (i.e., Ar + 6% CO2 and air + 6% CO2). For these cells, changes in xanthophylls are clear, whereby the magnitude of decreases in echinenone, canthaxanthin, and myxol 2′-fucoside and that of increases in 4-ketomyxol 2′-fucoside were all greater than those observed under the lowest C/N (air) treatment condition. This demonstrates the overall trend of a fine tuning of xanthophyll composition under differing environmental conditions and supports the important role of these carotenoids in extreme environments.

Discussion

Alginate-entrapped heterocystous cyanobacteria experience oxidative stress under H2 producing conditions

Our experimental work has shown that cyanobacterial cells entrapped within the alginate films and exposed to H2-photoproducing conditions encounter oxidative stress. Primarily, oxygen is evolved as a result of active water splitting during photosynthesis in the cyanobacteria entrapped in alginate films. Here, diffusion of oxygen to the gas phase through the alginate hydrogel matrix is not as efficient (Hassett 1996) as in suspension cultures. As a result, molecular oxygen trapped within the alginate matrix can directly interact with electron-transport cofactors producing ROS and thereby causing photodamage to the photosynthetic apparatus (Vass 2012) and/or inhibiting the PS II repair mechanism (Nishiyama et al. 2004). Such damage is particularly likely under conditions where rates of photosynthetic O2 evolution are increased, for example by higher light and CO2 levels (Kosourov et al. 2017).

In all three tested strains, oxidative stress was indeed evident in varying degrees toward the end of the experiment (450 h) as compared to the beginning of the experiment (0 h) (Fig. 3). The Western blots also showed a decrease in D1 protein levels in most samples of Anabaena strains throughout the experiment (Fig. 5). This was in contrast to Calothrix samples, where the decrease in D1 proteins was followed by a recovery (Fig. 5). As expected, degradation of D1 protein negatively affected the H2 photoproduction yields in ΔhupL and Calothrix cultures (Fig. 1a). In wild-type Anabaena, this effect was hindered by efficient H2 recycling via uptake hydrogenase. When photodamaged, D1 is degraded and re-synthesized in a process known as the PS II repair cycle (Mulo et al. 2012). Several studies have suggested that 1O2 might contribute to the photoinhibition of PS II by inhibiting its repair mechanism (Nishiyama et al. 2004; Inoue et al. 2011). Where there is excess 1O2, some cyanobacterial carotenoids may play a photoprotective role (Zhu et al. 2010; Kusama et al. 2015). We indeed observed a significant accumulation of myxoxanthophylls in immobilized cyanobacteria (Table 1). The presence of oxidative stress in cells was also confirmed by the accumulation of α-tocopherol (Fig. 4). However, the small differences observed in protein oxidation and α-tocopherol levels (Figs. 3 and 4) between different treatments suggest that the immobilization condition itself, rather than the condition of C/N imbalance, contributes most strongly to the oxidative stress observed.

The role of carotenoids in long-term H2 photoproduction

Cyanobacterial carotenoids comprise mostly carotenes and xanthophylls, the latter being the oxygenated derivatives of the former (Zakar et al. 2016). Carotenoids are an important part of the cyanobacterial photosynthetic machinery, particularly in the assembly and function of PSII. These pigments are known to undergo changes in quantity and composition when the cells are subjected to oxidative and other types of stress (Zakar et al. 2016) including the condition of the C/N imbalance (Kosourov et al. 2016). A similar trend was also observed in the current study (Table 1). It is clear that carotenoid composition was affected by both immobilization and long-term H2 photoproduction conditions. Carotenoids, like β-carotene and echinenone/3′-hydroxyechinenone, have major roles associated with the photosystems (Kerfeld 2004; Umena et al. 2011). Degradation of the reaction centers caused by stress conditions can result in the partial loss of the associated carotenoids in the cells, and we indeed observed a decrease in echinenone in almost all samples (Table 1). On the other hand, some carotenoids may contribute to the protection against photoinhibition and oxidative stress. Calothrix 336/3 showed a recovery of D1 protein by the end of the experiment and high H2 photoproduction yields, especially in air-treated films. At the same time, these samples showed a significant accumulation of nostoxanthin and two myxoxanthophylls: 2-hydroxymyxol 2′-methylpentoside and myxol 2′-methylpentoside (Table 1). The contribution of these carotenoids in photoprotection under H2-producing conditions is not clear at this point, but we suggest that the diverse carotenogenesis in Calothrix and accumulation of the above-mentioned xanthophylls may contribute at least partially to the resilience of this native strain to the N-deprived condition.

Uptake hydrogenase in heterocystous cyanobacteria indirectly contributes to photoprotection

In the alginate-entrapped ΔhupL films, there was a clear imbalance in oxygen evolution and H2 photoproduction yields under the air + 6% CO2 treatment, the condition where H2 photoproduction was lowest (Fig. 1). This result was despite lower PSII yield (Kosourov et al. 2014) and D1 levels (Fig. 5) and is likely due to the absence of uptake hydrogenase in ΔhupL strain, which removes excess oxygen in heterocysts. In cyanobacteria, uptake hydrogenase enzyme is responsible for removing excess oxygen in the heterocysts and furnishes the cells with ATP via oxyhydrogen reaction: 2H2 + O2 - > 2H2O (Bothe et al. 1977). By this process, the oxygen-sensitive nitrogenase enzyme remains free from inactivation. In ΔhupL, the lack of this enzyme and the compromised gaseous diffusion limitation of the alginate hydrogel matrix may together influence the accumulation of oxygen during active photosynthesis in the neighboring vegetative cells. The oxygen thus generated can seep into the heterocysts through the narrow terminal pores that connect heterocysts with vegetative cells (Walsby 2007). On exposure to oxygen, nitrogenase enzyme located in the heterocysts gets affected and impairs efficient nitrogen fixation during the periods of air treatments and thus fails to restore the photosynthetic apparatus, which is in agreement with a lower level of D1 and Chl a in the corresponding samples (Fig. 5 and Supp. Fig. S2). The photoprotective role of the uptake hydrogenase enzyme is also supported by the Western blots derived from the wild-type strain (Fig. 5), where D1 protein was mainly affected under the severe N-deficiency and C/N imbalance (Ar + 6% CO2-treated films).

Conclusion

In this work we have shown that cyanobacteria entrapped in alginate-based films and maintained under H2 photoproduction conditions employ different strategies to counteract the C/N imbalance and oxidative stress that they are exposed to over long-term use. Strain-specific strategies for dealing with different environmental conditions indicate the potential for fine-tuning cyanobacterial production platforms toward desired chemicals, as demonstrated here for xanthophylls. The native Calothrix strain employed in this study demonstrated particular resilience to both the C/N imbalance applied in standard H2-photoproducing conditions, which was associated with a distribution of excess carbon to the EPS and to the oxidative stress experienced during immobilization and consequent light exposure, where the elevated accumulation of α-tocopherol and total carotenoids were observed. While Calothrix demonstrated resilience to H2 photoproduction conditions, the uptake hydrogenase-deficient mutant (ΔhupL) of Anabaena was particularly susceptible to C/N imbalance, whereby cells had dramatically decreased levels of D1 protein and Chl a and had ceased producing H2 under all but the air treatment condition. To overcome the challenge of O2 accumulation inside the alginate films and better separate this stress from nutritional stress, we are currently exploring novel renewable and biodegradable materials with optimized, controllable porosity and mechanical stability, which may be more suitable for H2 photoproduction over the long term (Jämsä et al. 2018).